ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 8

Available online at w

journal homepage: www.elsevier .com/locate /watres

Influence of adsorption and anaerobic granular sludgecharacteristics on long chain fatty acids inhibition process

J. Palatsi a,b,*, R. Affes a,c, B. Fernandez a, M.A. Pereira b, M.M. Alves b, X. Flotats c

a IRTA, GIRO Joint Research Unit IRTA-UPC, Torre Marimon, E-08140 Caldes de Montbui, Barcelona, Spainb IBB e Institute for Biotechnology and Bioengineering, Centre of Biological Engineering, University of Minho, Campus de Gualtar,

4710-57 Braga, PortugalcDept. of Agrifood Engineering and Biotechnology, Universitat Politecnica de Catalunya (UPC), Parc Mediterrani de la Tecnologia, Edifici D-4,

E-08860 Castelldefels, Barcelona, Spain

a r t i c l e i n f o

Article history:

Received 11 May 2012

Received in revised form

4 July 2012

Accepted 6 July 2012

Available online 16 July 2012

Keywords:

Anaerobic granular sludge

LCFA inhibition

Adsorption isotherms

16S rDNA profiling

Microscopic techniques

* Corresponding author. IRTA, GIRO Joint ReTel.: þ34 938654350; fax: þ34 938650954.

E-mail address: jordi.palatsi@irta.cat (J. P0043-1354/$ e see front matter ª 2012 Elsevhttp://dx.doi.org/10.1016/j.watres.2012.07.008

a b s t r a c t

The impact of LCFA adsorption on the methanogenic activity was evaluated in batch

assays for two anaerobic granular sludges in the presence and absence of bentonite as

synthetic adsorbent. A clear inhibitory effect at an oleate (C18:1) concentration of

0.5 gC18:1 L�1 was observed for both sludges. Palmitate (C16:0) was confirmed to be the main

intermediate of C18:1 degradation in not adapted sludge and its accumulation was further

evidenced by fluorescence staining and microscopy techniques. LCFA inhibition could be

decreased by the addition of bentonite, reducing the lag-phase and accelerating the

kinetics of LCFA degradation, concluding in the importance of the adsorptive nature of the

LCFA inhibitory process. Granule morphology and molecular profiling of predominant

microorganisms revealed that biomass adaptation to LCFA could modify the intermediates

accumulation profiles and process rates.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction acetate (Ac), and finally to methane (CH4) by methanogenic

Anaerobic digestion is a highly sustainable waste treatment

process because it combines organic matter removal with

energy production in the form of biogas. The energy

yield depends, among other factors, on the organic matter

composition (generally defined as the ratio between lip-

idseproteinsecarbohydrates). In particular, lipids are inter-

esting substrates for the anaerobic digestion process due to

the high potential methane yield. Under anaerobic conditions,

lipids are initially hydrolyzed to glycerol and long chain

fatty acids (LCFA), which are further converted by acetogenic

bacteria (b-oxidation process) to hydrogen (H2) and

search Unit IRTA-UPC,

alatsi).ier Ltd. All rights reserve

archaea.

LCFAs are the main intermediate by-product of the lipid

degradation process, and their accumulation in anaerobic

digesters has been related with problems of sludge flotation,

biomass washout and inhibition of the microbial activity

(Rinzema et al., 1994). These LCFA inhibitory effects have been

associated to the interference with the electron transport

chain, impairment of the nutrient uptake, inhibition of

specific enzyme activities, or to the generation of toxic per-

oxidation and autooxidation products (Desbois and Smith,

2010). It has long been stated that adsorption of LCFA onto

the cell membrane is the main factor determining its

Torre Marimon, E-08140 Caldes de Montbui, Barcelona, Spain.

d.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 8 5269

biological toxicity (Galbraith and Miller, 1973). More recently,

Pereira et al. (2005) evidenced the importance of LCFA on

transport limitations at the level of the cell membrane, and on

how this affects the overall anaerobic digestion process. The

fact that LCFA inhibition is reversible suggests that it is

partially driven by LCFA adsorption onto the biomass and that

the methanogenic activity might thus be resumed once the

LCFA that have been accumulated onto the biomass are

progressively metabolized.

Oleate (C18:1), stearate (C18:0) and palmitate (C16:0) are the

major constituents in lipid-rich wastes and wastewaters

(Battimelli et al., 2010; Valladao et al., 2011). The inhibitory

effect of a specific LCFA has been defined as concentration

dependent, but it is also related to the LCFA chain length and

degree of saturation (Lalman and Bagley, 2001). Differences on

cell membrane properties between bacteria and archaea also

play an important role on LCFA toxicity (Zheng et al., 2005).

C16:0 has been proposed to be the main intermediate and the

key inhibitory specie during the anaerobic degradation of

C18:1 via the b-oxidation process (Pereira et al., 2002).

Recently, several studies have published valuable information

dealing with the diversity of saturated/unsaturated LCFA

degraders and syntrophic methanogens interactions (Sousa

et al., 2007).

Different strategies have been studied to prevent or to

recover lipid/LCFA inhibited systems. The solubilization of

lipid waste via saponification (Battimelli et al., 2010) or enzy-

matic pre-treatments (Valladao et al., 2011), the application of

feeding procedures based on sequential LCFA accumulation-

degradation steps (Cavaleiro et al., 2009), the addition of

easily degradable co-substrates (Kuang et al., 2006) or the

addition of adsorbents as recovery agents (Palatsi et al., 2009)

have been proposed as possible strategies to limit the inhibi-

tory effects.

Despite the extensive literature references about LCFA

inhibition, only few studies focus on how important the

adsorption process during the LCFA degradation is (Hwu et al.,

1998; Pereira et al., 2005). The aim of the present study is to

monitor the LCFA adsorption over granular sludge and to

evaluate its impact on process inhibition, bymeans of: specific

batch tests (monitoring LCFA evolution), addition of synthetic

adsorbents (bentonite), molecular profiling of the predomi-

nant microorganisms and fluorescence staining microscopy

imaging.

2. Material and methods

2.1. Physicochemical and biological characterization ofbiomass

Two different anaerobic granular sludges were used in

present experiments, sampled from industrial beverage

wastewater UASB reactors: sludge-A from a beer brewery

(A Coruna, Spain) and sludge-B from a fruit juice processing

industry (Lleida, Spain). Both sludges were characterized in

terms of granular morphology, methanogenic activity and

bacterial community structure.

The characterization of the granular morphology of both

sludges (20 samples, containing more than 1200 granules per

sample or >0.2 g per sample) was performed analyzing digi-

talized images (768 � 574 pixel size, 256 grey levels) with the

Analyze Particle Tool of ImageJ package software (National

Institutes of Health, USA). Imageswere binarized and particles

sizes were evaluated by equivalent diameter and specific

surface (cm2 g�1VSS), calculated from the particles projected

area, according to Pereira et al. (2003).

Methanogenic activity tests (SMA) of both sludges on

acetate (Ac) and hydrogen (H2/CO2) were performed in batch

anaerobic vials (120mL total volume vials with 50mL ofmedia

working volume) at mesophilic temperature (35 �C), as

described by Angelidaki et al. (2009).

The original microbial community structure was depicted

by molecular profiling. The total DNA of biomass samples of

both sludges was extracted and bacterial 16S rDNA gene

fragments were amplified by polymerase chain reaction

procedure (PCR). Amplicons were subsequently resolved by

denaturing gradient gel electrophoresis (DGGE) analysis.

Relevant DGGE bands were excised, reamplified by PCR and

sequenced as described in Palatsi et al. (2010). Sequences were

edited using the BioEdit software package v.7.0.9 (Ibis Biosci-

ences, USA), compared against the NCBI genomic database

with the BLAST search alignment tool (Altschul et al., 1990),

and related to phylogenetic groups by using the RDP Naive

Bayesian Classifier (Wang et al., 2007). Obtained nucleotide

sequences have been deposited in the GenBank database

under accession numbers HM100129eHM100135.

2.2. LCFA adsorption isotherms on bentonite andgranular sludge

The adsorption isotherms for oleate on bentonite and on

anaerobic granular biomass (sludge-A) were determined in

batch experiments (1 L vials with 500 mL of media working

volume). Media was composed by demineralized water,

sodium bicarbonate (1 gNaHCO3g�1

CODadded) and bentonite or

inactivated anaerobic granular sludge as sorbents. In order to

differentiate LCFA adsorption from biological degradation,

biomass was previously inactivated by incubation (2 h at 4 �C)with formaldehyde solution (4% v/v). This protocol was

selected to minimize cell wall damage produced by other

protocols (Ning et al., 1996). Sodium oleate at a concentration

range from 0.5 to 5.5 gC18:1 L�1 was added as LCFA sorbate

model. Vials were maintained at mesophilic (35 �C) anaerobicconditions under continuous shaking (150 rpm).

Liquid samples were withdrawn periodically from the vials

to monitor the soluble C18:1 concentration (C;

mgC18:1soluble L�1). These measurements were fitted to a time

course asymptotical exponential decay curve (C ¼ Ce þ ae�ßt)

to determine the equilibrium C18:1 concentration in the liquid

phase (Ce; mgC18:1soluble L�1). Obtained Ce values, for each

tested concentration, were used to fit a Langmuir isotherm

model (Eq. (1)):

Cad ¼ CmaxKCe

1þ KCe(1)

where Cad is the equilibrium amount of sorbate per unit of

sorbent (mgC18:1adsorbed g�1TSadsorbent), Cmax is the maximum

sorptivity of C18:1 on sorbent (mgC18:1adsorbed g�1TSadsorbent)

and K is the Langmuir equilibrium constant (L mg�1C18:1soluble).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 85270

Sorbent concentrations were expressed in total solids (TS)

units to compare biomass and bentonite adsorption capacity.

2.3. Selection of LCFA inhibitory concentration

The inhibitory effect of LCFA over anaerobic granular sludges

(sludge-A and sludge-B) was assessed in batch assays.

Increasing concentrations of C18:1 from 0.1 to 1.5 gC18:1 L�1

were tested in 120 mL glass vials (50 mL of working

volume) containing 5 gVSS L�1 of granular sludge and

1 gNaHCO3g�1

CODadded, under strict mesophilic (35 �C) anaerobicconditions and continuous shaking (150 rpm). Those tests

were carried out to determinate the amount of C18:1 that

causes a clear and long lasting inhibition on biogas production

over the selected inoculum. Specific methane rates were

determined by the initial slope of the accumulated CH4

production curve per biomass content unit (mLCH4 g�1

VSS d�1),

to better compare different sludges responses.

2.4. Influence of the adsorptive process over LCFAinhibition

Three different experiments with bentonite addition were

tested in batch assays in order to evaluate and quantify the

effect of the LCFA-adsorption over LCFA inhibitory process.

Main experimental set-up, including timing for sludge

exhaustion, methanogenic activation and oleate or bentonite

addition was summarized in Table 1.

The first experiment (E1) was performed with sludge-A

(Ta_1 vials in Table 1) and consisted in testing the possibility

of prevention granular sludge inhibition by means of a prelim-

inary bentonite addition into vials (day �2), prior the LCFA

inhibitory pulse (day þ1). Contrary, the second experiment

(E2), also performed with sludge-A (Ta_2 vials in Table 1), was

focused on testing the possibility of recovery the degradation

capacity of inhibited granular sludge (LCFA pulse in dayþ1) by

means of bentonite post-addition (day þ2). As sludge-A was

sampled months before the present experiments (stored at

4 �C), a biomass activation stepwith an acetate pulse (day 0, in

Table 1) was included in E1 and E2 experiments to check the

sludge methanogenic activity.

The third experiment (E3) was performedwith sludge-B (Tb

vials in Table 1) and was intended to capture LCFAs by incu-

bating a higher concentration of bentonite with oleate (day-4),

Table 1 e Experimental set-up of batch experiments to evalua

Experiment E1 and E2 Day �2 (sludge exhaustion)

Ta_1 (E1) 0.5 gBentonite L�1 þ 5 gVSS(A) L

�1 þ 3 gNaHCO3L�1

Ta_2 (E2) 5 gVSS(A) L�1 þ 3 gNaHCO3

L�1

Ca 5 gVSS(A) L�1 þ 3 gNaHCO3

L�1

BLa 5 gVSS(A) L�1 þ 3 gNaHCO3

L�1

Experiment E3 Day -4 (LCFA-bentonite adsorption)

Tb (E3) 5.0 gBentonite L�1 þ 0.5 gC18:1 L

�1 þ 3 gNaHCO3L�1

Cb 0.5 gC18:1 L�1 þ 3 gNaHCO3

L�1

BLb 3 gNaHCO3L�1

to force exclusively the LCFA adsorption over bentonite, prior

to vials inoculation with granular sludge (day 0). Since sludge-

B arises from a steady state large-scale reactor, and was

sampled immediately before the present assay, the sludge

was previously exhausted (day-2) but the activation step with

acetate was omitted.

In all experiments (Table 1), vials were previously buffered

ð3 gNaHCO3L�1Þ and the initial biomass and oleate concentra-

tions were set constants (5 gVSS L�1and 0.5 gC18:1 L�1, respec-

tively). Control vials (Ca and Cb), with LCFA and biomass but

without bentonite, and blank vials (BLa and BLb) with only

biomass, were also run for all the experiments (Table 1).

Incubations were carried out at 35 �C under shaking (150 rpm)

and anaerobic conditions. Each treatment was performed in

triplicate for biogas analysis (CH4), and 12 vials per treatment

were withdrawn during incubations in order to determine

long chain fatty acids present in the solid and liquid phase

(LCFAS and LCFAL, respectively), and the volatile fatty acids

(VFA) profile. All parameters were expressed in COD units to

better follow the mass balance.

To quantify the adsorption effect of bentonita addition,

a Gompertz equation (Eq. (2)) was fitted to the methane

production profiles in E2 experiments:

P ¼ Pmax$exp

�� exp

�rm$e

Pmax$ðt� lÞ þ 1

��(2)

where P is the accumulated methane production

(mg CODCH4 L�1), which is expressed as a time t function (day),

Pmax is the methane production potential (mg CODCH4 L�1), rmis the maximum methane production rate

(mg CODCH4L�1 day�1) and l is the lag phase period on biogas

production (day).

2.5. Microscopy observations

For microscopy observation, intermediate (day þ10) samples

of granular sludge-A vials (Table 1), free (BLa) and submitted to

LCFA pulse (Ca), were fixated (4% v/v formaldehyde in PBS) and

embedded in Tissue-Tek OCT (optimal cutting temperature)

compound (Sakura Finetek), as described in Batstone et al.

(2004). Granule sections of 10 mm were obtained with a cryo-

stat (CM 1900 Leica, Germany), operated with a knife

temperature of �20 �C and a cabinet temperature of �18 �C.Slides were stained with multiple fluorochrome dyes. DAPI

te LCFA adsorption effect.

Day 0 (activation) Day þ1 (inhibition) Day þ2 (recovery)

þ30 mMAc þ0.5 gC18:1 L�1 e

þ30 mMAc þ0.5 gC18:1 L�1 þ0.5 gBentonite L

�1

þ30 mMAc þ0.5 gC18:1 L�1 e

þ30 mMAc e e

Day �2 (sludge exhaustion) Day 0 (inoculation)

0.5 Lsludge B under anaerobic conditions þ5 gVSS(B) L�1

þ5 gVSS(B) L�1

þ5 gVSS(B) L�1

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 8 5271

(4,6-diamidino-2-phenylindole, Sigma) was used as probe for

biomass or total cells, while Nile Red (9-diethylamino-5H-

benzo[a]phenoxazine-5-one, Sigma) was selected as probe for

hydrophobicity sites (lipids), based on Diaz et al. (2008).

Images were obtained with a BX51 (Olympus, Japan) fluores-

cence light microscope (FLM). Observation settings were

Ex365-370/B400/LP421 and Ex530-550/B570/LP591 for blue

(DAPI) and red (Nile Red) channels, respectively. Merged

images were obtained using ImageJ (National Institutes of

Health, USA) package software.

2.6. Analytical methods

Sodium oleate powder salt (Riedel-de Haen/SigmaeAldrich;

82% C18:1/LCFA) was selected as LCFA substrate model due to

its high solubility and itsmajor presence on fatty wastewaters

(Hwu et al., 1998). Analytical grade powder bentonite (Al2O3

4SiO2 H2O, SigmaeAldrich) was selected as the synthetic

adsorbent.

Total solids (TS), volatile solids (VS), volatile suspended

solids (VSS) and pH were determined according to Standard

Methods (APHA, 1995). Biogas and methane (CH4) production

was monitored by pressure transducer and gas chromatog-

raphy techniques (FID and TCD), as described elsewhere

(Angelidaki et al., 2009). Volatile fatty acids (VFA)eacetate

(Ac), propionate (Pro), iso and n-butyrate (Bu) and iso

and n-valerate (Va) were determined after sample acidifica-

tion with a CP-3800 gas chromatograph (Varian, USA), fitted

with Tecknokroma TRB-FFAP capillary column

(30 m � 0.32 mm � 0.25 mm) and FID detection (Palatsi et al.,

2009). Long chain fatty acids (LCFA) e laurate (C12:0), myr-

istate (C14:0), palmitate (C16:0), palmitoleate (C16:1), stearate

(C18:0) and oleate (C18:1) e were determined as fatty acids

methyl esters (FAME), based on direct LCFA methilation-

extraction procedure, with a CP-3800 gas chromatograph

(Varian, USA), fittedwith the capillary columnVarian CP-Sil 88

(50m � 0.25mm � 0.2 mm) and a FID detector, according to

Palatsi et al. (2009). Obtained experimental samples were

centrifuged (4500 rpm 10 min), to differentiate between solid

and liquid phase LCFA content (LCFAS or LCFAL, respectively).

Fig. 1 e Estimated adsorption isotherms, using a Lagmuir

isotherm model, for bentonite and inactivated sludge-A,

and comparison with available literature values. Markers

represent experimental data while lines represent data

fitting to isotherm models. Coefficients of determination

for data fitting (R2) are also indicated in the figure legend.

3. Results and discussion

3.1. LCFA adsorption isotherms on bentonite andgranular sludge

Bentonite has been considered a good adsorbent for organic

species because of its high specific surface area, high porosity

and surface activity. Scanning electron microscopy (SEM) and

FT-IR spectrometric analysis evidenced the porous structure of

the claymineral and thepossible interactionof LCFAmolecules

with the silanol (SiO2) groups of bentonite (Demirbas et al.,

2006). A Langmuir model was fitted against the experimental

data and the obtained parameters were compared with the

literature (Fig. 1). The obtained oleate maximum sorptivity

value (Cmax) on sludge-Awas373.78mgC18:1adsorbed g�1

TSadsorbent

and the equilibriumconstant (K ) was 1.7$10�3 Lmg�1C18:1soluble.

These isotherm parameters were lower than those previously

reported by Hwu et al. (1998), using a Freundlich isotherm

model and thermally inactivated granular sludge. With

respect to bentonite, the estimated Cmax and K

values were 901.76 mgC18:1adsorbed g�1TSadsorbent and

1.23$10�2 L mg�1C18:1soluble, respectively, which evidenced the

relatively high adsorption capacity of the clay based mineral

sorbent compared with biomass (sludge-A). These values were

comparable to those obtained by Mouneimne et al. (2004) with

bentonite suspensions and solid fatty wastes.

Assuming that adsorption can be described as a physical

phenomenon dependent on the sorbent surface and sorbate

concentration (Ning et al., 1996; Hwu et al., 1998), the higher

specific surface area of clay minerals like bentonite (Raposo

et al., 2004) in relation to the anaerobic granular sludge

(further reported), explains the higher adsorption capacity of

the former. These results show that bentonite outcompetes

granular biomass in terms of sorptive capacity towards LCFA

and, thus, its addition as a synthetic adsorbent in anaerobic

digesters might influence dynamics of LCFA-adsorption-

inhibition process.

3.2. LCFA inhibitory concentration

LCFA degradation rates by sludge-A and sludge-B were esti-

mated, according to Material and Methods section, using the

initial slope of the specific methane production rate at

different initial LCFA concentrations. Exponential decay

curves were fitted to the obtained degradation rates (Fig. 2).

A concentration of 0.5 gC18:1 L�1 (aprox. 1.5 gCOD L�1) caused

a decrease on themethane production rate higher than 50% in

both sludges (sludge-A and sludge-B), according to Fig. 2. This

value was selected as the reference oleate concentration in

subsequent batch experiments (Table 1). It is interesting to

observe that sludge-B displayed a higher degradation capacity

of oleate in relation to sludge-A, at least up to concentrations

of 1 gC18:1 L�1 (Fig. 2). The different resistances of sludge-A and

sludge-B to C18:1 inhibition will be discussed later, consid-

ering the results of biomass characterization (granules

morphology, methanogenic activity and molecular profiling).

Fig. 2 e Effect of tested initial LCFA (C18:1) concentration

(expressed as gC18:1 LL1) on the initial specific methane

production rate (mLCH4 gL1VSS dL1) of granular sludge-A

and sludge-B. Markers represent experimental values

while lines represent the fitting to an exponential decay

curve. Coefficients of determination for data fitting (R2) are

also indicated in the figure legend.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 85272

3.3. Influence of the adsorptive process on LCFAinhibition

Based on the previous results (adsorption isotherms and LCFA

inhibitory concentration determination) different experi-

ments were tested to study the effect of adsorptive process on

LCFA inhibition and to quantify the LCFA adsorption-

inhibitory effect by applying bentonite additions, as

described previously in the Materials and Methods section

(Table 1).

3.3.1. Addition of bentonite prior or after LCFA inhibition(E1 and E2)The time course evolution of the LCFAS and LCFAL uptake,

acetate transient accumulation, and methane production

were monitored in E1 and E2, and showed in Fig. 3.

Considering that the oleate pulse was added from an

aqueous stock solution, it was assumed that this LCFA was

exclusively contained in the liquid phase (Fig. 3b and d). The

adsorption of LCFA from the liquid media to solid phase was

quite fast in all treatments, including the control vials (Ca) with

no bentonite (Fig. 3a and c). Oleate uptake from the solid phase

was followed by a clear accumulation of palmitate, primarily

on the solid phase as well, as shown in Fig. 3c. No significant

amounts of stearate nor palmitoleate were detected in liquid

or solid phase samples (data not shown), which is consistent

with the hypothesis of hydrogenation of unsaturated LCFA

prior to the b-oxidation process (Lalman and Bagley, 2001).

The metabolism of LCFA resulted in a transient accumu-

lation of acetate in themedium (Fig. 3e), which lasted until the

complete degradation of palmitic acid. No significant amounts

of other VFA intermediates (valerate, butyrate or propionate)

were detected (data not shown), which is in agreement with

others studies (Cavaleiro et al., 2009). The inhibition caused by

the LCFA pulse also resulted in an immediate halt onmethane

production. However, this process was reversible as methane

formation resumed after 30e35 days (Fig. 3f), when the

remaining LCFA were completely depleted (Fig. 3a and c).

No significant differences were observed in LCFA degrada-

tion, accumulation of acetate, and methane production,

between bentonite (Ta_1) and control (Ca) vials in experiments

E1 (Fig. 3), indicating that bentonite had little effect on pre-

venting or modifying LCFA inhibition. A possible explanation

for this unexpected phenomenon in E1 is the potential inter-

action between bentonite and biomass prior to the oleate pulse

(4 days contact, Table 1). Clay minerals, like bentonite, have

been largely used as support materials for the growth of

anaerobic biofilms, due to their biomass adsorption capacity

(Arnaiz et al., 2006). Furthermore, the spatial distribution and

probability of bentonite-LCFA or biomass-LCFA adsorption

occurrence, which is a function of LCFA concentration and

particle density (0.5 gBentonite L�1 vs 5 gVSSbiomass L�1), might

have played an important role in the results. Mouneimne et al.

(2004) considered this factor in thecalculationof theadsorption

isotherms for multiple species systems (biomass-clay mineral-

LCFA). Consequently, if biomass was attached to bentonite

particles prior to LCFA pulse, the preventing effect expected in

experiment E1 might have been severely hampered as the free

adsorption sites available for LCFA in bentonite surface might

have experienced a strong reduction (see scheme a in Fig. 4).

On the other hand, relatively few differences were

observed in the LCFA degradation profile and methane

production rates, between bentonite (Ta_2) and control vials

(Ca) in the experiment E2, which resulted in a slightly faster

acetate consumption andmethane production rate (Fig. 3). If it

is assumed that LCFA b-oxidation process was performed over

the granule cell wall surface (only Ac and H2 delivered to the

media, but not other LCFA intermediates), the addition of

bentonite would affect only the initially (day þ2, Table 1)

added LCFA concentration not yet adsorbed onto biomass (low

according to Fig. 3). Consequently, the recovering effect of E2

would be poor (see scheme b in Fig. 4).

In order to confirm that the measured accumulation of

palmitic acid on the solid phase (LCFAs) is effectively adsorbed

onto the granule surface (no b-oxidation intermediates deliv-

ered to themedia) causing biomass inhibition, rather than just

precipitating in the bulk media, granules from the control (Ca)

and blank vials (BLa) were taken at day þ10, when most of the

oleate had already been consumed and maximum palmitate

levels were detected in the solid phase (Fig. 3c). Granules from

Ca (incubations with biomass and oleate) and BLa (only

biomass for endogenous activity), vials were sectioned,

stained with DAPI-Nile Red, and examined under fluorescent

light microscopy (FLM). Fig. 5 shows an example of the

appearance under FLM of BLa (Fig. 5aec) and Ca (Fig. 5def)

stained sections.

Nile Red is a hydrophobicity indicator but with this stain it

is not possible to differentiate between the natural phospho-

lipids of the cell membrane and the adsorbed LCFA (Diaz et al.,

2008). Yet, the more intense Nile Red signal observed in the

outer layer of the Ca granules (white arrow in Fig. 5f) in rela-

tion to the BLa further supports the fact that palmitate formed

upon the metabolism of oleate is adsorbed on the granule

surface, with the subsequent potential implications in

membrane transport limitations or in other LCFA inhibitory

effect.

Fig. 3 e Comparison between E1 and E2 batch tests in terms of LCFA concentration in the solid phase (a and c) and in the

liquid phase (b and d), acetate profile (e) and methane formation (f), for bentonite (TA_1 or TA_2), control (CA) and blank (BLA)

vials. All parameters are expressed in COD equivalent concentration units (mgCOD LL1). The circles indicate the initial

estimated concentration from LCFA pulse added into the vials.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 8 5273

3.3.2. Addition of bentonite for LCFA capturing (E3)To confirm the implications of LCFA adsorption over process

inhibition, a different new experiment (E3) was designed with

sludge-B. In this case, bentonite was also incubated with

Fig. 4 e Scheme of the bentonite addition effect onto LCFA

adsorption-inhibition process in E1, E2 and E3

experiments.

oleate during 4 days prior to biomass addition (Table 1), in

order to promote LCFA-bentonite adsorption or LCFA capturing

for the enhanced protection of biomass from LCFA

adsorption-inhibition. A higher content of adsorbent

(5 gbentonite L�1) was added to have a more even biomass/

bentonite particle distribution ratio, as previously reported by

Mouneimne et al. (2004). The time course evolution of the

LCFAS and LCFAL uptake, acetate transient accumulation, and

methane productionweremonitored in E3, according to Fig. 6.

The plot scales were maintained to that of the previous

experiments (Fig. 3), in order to facilitate the direct compar-

ison of the obtained results.

Since in E3 oleate was incubated during 4 days in a buff-

ered medium with and without bentonite (Table 1), at the

time of biomass addition (day 0) it was considered that all

oleate was solubilized in Cb vials and adsorbed on Tb vials

(Fig. 6a and b). From the evolution of the oleate biodegrada-

tion pattern, clear differences were observed among treat-

ments. Oleate biodegradation rates were higher when

bentonite was added (Fig. 6a). A similar positive effect has

Fig. 5 e FLM images of DAPI staining (a and d), Nile Red staining (b and e) and merged emissions (c and f), of BLA and CA

sectioned granules, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to

the web version of this article.)

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 85274

been described on sepiolite addition during anaerobic degra-

dation of oleate (Cavaleiro et al., 2010). This phenomenon was

then claimed to be the result of a decrease in the LCFA

inhibitory effect because of the physical adsorption, but also

to an improved metabolite transfer between acetogens and

methanogens due to the higher proximity between those

groups promoted by biofilm formation. This effect was not

observed in the experiment E1, where biomass-bentonite

adsorption was promoted (Fig. 3). Furthermore, in the

experiment E3, oleate degradation did not produce the ex-

pected accumulation of palmitate in the solid phase, neither

in the treatment vials nor in the controls (Fig. 6c), which

points to the fact that sludge-B was probably more adapted to

the metabolism of LCFA, as already indicated by the

comparatively higher oleate biodegradation rate (Fig. 2).

Cavaleiro et al. (2009) reported transient palmitate accumu-

lation in reactors fed with oleate-based influents but, because

of biomass adaptation, palmitate was not longer detected

after two or three cycles of LCFA feeding. To clarify these

aspects further biomass (sludge-A and sludge-B) character-

ization (granule morphology, methanogenic activity and

bacterial community structure) was performed and further

discussed in the next paragraphs, to explain their different

LCFA biodegradation patterns. It must be noticed as well that

sludge-B consisted of fresh granules collected from an oper-

ative UASB reactor from a fruit juice processing industry and,

despite the preliminary starvation step, it still conserved

a relatively high residual organic matter when compared to

sludge-A (BLb vs BLa, methane production in Figs. 3 and 6).

The presence of easily biodegradable co-substrates, such as

glucose and cysteine, which are likely to be present in higher

amounts on sludge-B, has been found to enhance the LCFA

degradation rates (Kuang et al., 2006). Despite the intrinsic

differences on the assayed sludges, the LCFA adsorptive

action of bentonite in experiment E3, capturing LCFA,

confirmed the implications of the LCFA-adsorptive process

onto biomass, and confirms the adsorbent addition to be

a reliable approach to improve the process robustness, by

intervening on the kinetics of the LCFA adsorption and bio-

logical inhibition process (Fig. 4c). In particular, the acetate

accumulation was prevented (Fig. 6e) and methane produc-

tion was improved (Fig. 6f) upon bentonite addition. A Gom-

pertz equation (Eq. (2)) was fitted to the methane production

profile on both, treatment (Tb) and control (Cb) vials. as

described in Fig. 6f. For an equivalent estimated Pmax in Cb

and Tb vials (2569.4 � 106.6 and 2445.6 � 32.9 mg CODCH4 L�1,

respectively), the lag phase for bentonite vials (Tb) was

significantly decreased from 6.2 � 0.8 to 3.0 � 0.7 days (51.3%)

whilst the maximummethane production rate was improved,

from 130.5 � 12.2 to 187.1 � 10.3 mg CODCH4 L�1 day�1 (43.3%).

Thus, it can be concluded that the LCFA-adsorptive process

partially explain the LCFA inhibition process (nutrient trans-

port limitation) with a clear effect on the overall degradation

process, but there are still some initial inhibitory or toxic

effect (cell or membrane damage) that must be also consid-

ered. Further complementary studies, which include mathe-

matical modeling of the experimental data, are underway and

will help to elucidate the relevance of specific factors that are

associated to membrane damage.

3.4. Influence of sludge characteristics on LCFAinhibition

Considering that adsorption is a surface-related phenom-

enon, sludge specific area might be related to the LCFA

inhibitory effect as well, as earlier proposed (Hwu et al.,

1998). Therefore, the specific surface areas (cm2 g�1VSS) of

both granular sludges were estimated by image analysis

(Table 2). The obtained values on the specific area of sludges-

A and sludge-B were quite similar and did not explain the

differences on the LCFA inhibitory effect. On the other hand,

LCFA-degrading microorganisms are known to be proton

Fig. 6 e E3 batch tests in terms of LCFA concentration in the solid phase (a and c) and in the liquid phase (b and d), acetate

profile (e) and methane formation (f), for bentonite (TB), control (CB) and blank (BLB) vials. All parameters are expressed in

COD equivalent concentration units (mgCOD LL1). The circles indicate the initial estimated concentration from LCFA pulse

added into the vials.

Table 2 e Summary of granular biomass characterization(mean values and standard deviation).

Parameter Sludge-A Sludge-B

Sludge biomass content

(gVSS L�1)

8.81 � 0.13 8.87 � 0.02

Mean size as equivalent

diameter (mm)

1.96 � 0.72 2.04 � 0.93

Specific surface area

(cm2 g�1VSS)

620.50 � 41.57 540.13 � 58.65

Hydrogenotrophic

SMA (mLCH4 g�1

VSS d�1)

179.66 � 1.66 88.26 � 9.09

Acetoclastic SMA

(mLCH4 g�1

VSS d�1)

128.93 � 14.81 69.69 � 1.97

C18:1 initial content

(mg g�1 TSbiomassa)

<d.l 10.5 � 0.7

C16:0 initial content

(mg g�1 TSbiomassa)

16.5 � 2.1 105.0 � 2.8

SMA-specific methanogenic activity.

a Lyophilized biomass. <d.l ¼ under detection limit.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 8 5275

reducing bacteria, which interact syntrophically with H2-

utilizing and acetoclastic methanogens (Schink, 1997;

Lalman and Bagley, 2001). Consequently, the balance

between acetogenic bacteria and archaeal communities

plays a central role in the LCFA metabolism, influencing the

overall process rate (Pereira et al., 2002). Nevertheless, since

sludge-B had a lower methanogenic activity (both aceto-

clastic and hydrogenotrophic activity) than sludge-A, it is not

possible to conclude that differences concerning LCFA

biomass tolerance are directly related to the methanogenic

activity (Table 2).

Acetogenic and/or b-oxidizing bacteria can also influence

sludge behavior in relation to LCFA tolerance and metabo-

lism. A considerable increase in the b-oxidative activity of

anaerobic sludge has been observed upon LCFA pulses

(Pereira et al., 2005; Palatsi et al., 2010). Significant differences

in the microbial community structure of bacterial pop-

ulations from initial biomass samples (sludge-A and sludge-

B) were observed upon DGGE profiling of PCR amplified

Fig. 7 e DGGEprofilesof eubacterial 16S rDNAamplified from

sludge-A and sludge-B. A standard ladder (L) has been used

at both gel ends in order to check the DNA migration

homogeneity. Successfully excised and sequenced bands

have been named with lower-case letters.

Table 3 e DGGE bands: designations and accession numbers forganisms.

Band Accession no aClosest organism in Gedatabase (access no.) Ge

closest cultured/type strain

Band a HM1000129 Uncultured (AB266993)

Geobacter hephaestius strain VeEG4Tc

Band b HM1000130 Uncultured (CU923168)

Levilinea saccharolytica strain KIBI-1T

Band c HM1000131 Uncultured (DQ443992)

Aminomonas paucivorans stain GLU-3

Band d HM1000132 Uncultured (CU918694)

Petrimonas sulfuriphila stain BN3 (AY

Band e HM1000133 Uncultured (EU937734)

Brooklawnia cerclae strain BL-34Tc (D

Band f HM1000134 Uncultured (AB175367)

Cellulophaga tyrosinoxydans strain EM

Band g HM1000135 Uncultured (AB267007)

Paludibacter propionicigenes strain W

a Sequences were matched with the closest relative from the GenBank d

b Sequences were related to phylogenetic groups by using the RDP Naiv

c T e Type strain.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 85276

bacterial 16S rDNA ribotypes (Fig. 7). Up to 7 bands from the

obtained profiles were successfully excised, reamplified and

subsequently sequenced (Table 3). BLAST sequence

comparisons against the NCBI genomic database resulted in

close matches with uncultured ribotypes in most of cases,

previously described in anaerobic granular reactors (Narihiro

et al., 2009; Riviere et al., 2009). Several phylogenetic groups

were identified, mainly related with fermenting bacteria and

acetate or propionate oxidative microorganisms, but none of

themwas directly related to specific LCFA degrading bacteria.

Considering the origin of the tested sludges, from industrial

plants where lipids might only be present in residual

amounts, the absence of bacterial species related to LCFA

metabolism among the dominant fraction of the bacterial

community is thus not surprising. Consequently, the initial

number of bacteria responsible for LCFA degradation may

not be high enough for yielding visible bands in a DGGE

profile. Other molecular high throughput sequencing tech-

niques, such as pyrosequencing, might give a deeper insight

in relation to the presence and biodiversity of b-oxidizing

bacteria. Nevertheless, the obtained results clearly reveal

that the origin of the biomass is relevant to the LCFA

metabolism. Sludge-A was obtained from a brewing industry,

where its wastewaters are expected to be free of lipids.

Conversely, sludge-B was obtained from a fruit juice facility,

and the wastewater could have been in contact with

higher lipid content from fruit peels or press liquor wastes.

This fact would be in agreement with the analysis of the

initial LCFA profile performed on both sludges (Table 2), in

which a significantly higher amount of LCFA was found in

sludge-B compared to sludge-A. Further complementary

studies, which include mathematical modeling of the

experimental data, are underway and will help to elucidate

the relevance of the specific factors that are associated to the

biomass.

or the band sequences and levels of similarity to related

nBanknBank(access no)

Similarity % bPhylogenetic group

100.0 Geobacter

(AY737507) 97.8

100.0 Anaerolineaceaec (AB109439) 92.1

99.8 Synergistaceae

(AF072581) 89.9

100.0 Petrimonas sulfuriphila

570690) 100.0

99.3 Brooklawnia

Q196625) 97.3

100.0 Bacterioides

41Tc (EU443205) 91.1

98.0 Paludibacter

B4Tc (AB078842) 91.5

atabase.

e Bayesian Classifier (Wang et al., 2007).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 8 5277

4. Conclusions

Batch incubations of methanogenic biomass from two

different origins have clearly shown the inhibitory effect of

oleate at concentrations above 0.5 gC18:1 L�1. Nevertheless, this

process was reversible and methanogenic activity recovered,

showing that partial b-oxidation was a non-limiting step.

Palmitate was confirmed as the main intermediate in less

adapted biomass and the main LCFA inhibitory species. Solid

and liquid LCFA characterization and fluorescence staining

microscopy inspection of the biomass, performed once all

oleate was consumed from the liquid phase, demonstrated

that palmitate was mostly adsorbed on granules surface.

Oleate adsorption over anaerobic granular sludge and over

bentonite was characterized by means of Langmuir

isotherms, which revealed the higher adsorption capacity of

bentonite. Batch experiments designed to quantify the

importance of the LCFA adsorption process on biological

inhibition, by means of bentonite addition, proved that the

protection of granule surface from LCFA-rich streams is

a reliable approach to improve the system robustness, by

affecting LCFA-biomass adsorption dynamics. Nevertheless,

there is part of the inhibitory effect that still must be related

with a LCFA toxic effect. Yet, biomass adaptation to LCFA is

equally important for the anaerobic treatment of lipids.

Acknowledgments

This work was funded by the Spanish Ministry of Science and

Innovation (projects ENE 2007-65850 and CTM 2010-18212),

and was partially supported by a grand from the Department

of Universities, Research and Media Society of Catalonia

Government (BE-DGR 2008 BE1 00261). We would like to thank

Lucia Neves, Ana Nicolau, Madalena Vieira, and Ana Julia

Cavaleiro, from University of Minho, for their assistance in

microscopic observations and analytical methods. We also

thank Miriam Guivernau (IRTA) for assistance in the PCR-

DGGE profiling and ribotype sequencing. We are also grateful

to David Bedoya (MWH) and Francesc Prenafeta (IRTA) for the

revision and critical reading of the manuscript.

r e f e r e n c e s

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990.Basic local alignment search tool. Journal Molecular Biology215, 403e410.

Angelidaki, I., Alves, M., Bolzonella, D., Borzacconi, L., Campos, L.,Guwy, A., Jenicek, P., Kalyuzhnui, S., van Lier, J., 2009. Definingthe biomethane potential (BMP) of solid organic wastes andenergy crops: a proposed protocol for batch assays. WaterScience and Technology 59 (5), 927e934.

APHA, AWWA, WEF, 1995. Standard Methods for the Examinationof Water and Wastewater, nineteenth ed. American PublicHealth Association, American Water Works Association,Water Environment Federation, Washington DC, USA.

Arnaiz, C., Gutierrez, J.C., Lebrato, J., 2006. Support materialselection for anaerobic fluidized bed reactor by phospholipidanalysis. Biochemical Engineering Journal 27, 240e245.

Batstone, D.J., Keller, J., Blackall, L.L., 2004. The influence ofsubstrate kinetics on the microbial community structurein granular anaerobic biomass. Water Research 38,1390e1404.

Battimelli, A., Torrijos, M., Moletta, R., Delgenes, J.P., 2010.Slaughterhouse fatty waste saponification to increase biogasyield. Bioresource Technology 100, 3388e3393.

Cavaleiro, A.J., Salvador, A.F., Alves, J.I., Alves, M., 2009.Continuous high rate anaerobic treatment of oleica cid basedwastewater is possible after a step feeding start-up.Environmental Science and Technology 43, 2931e2936.

Cavaleiro, A.J., Sousa, D.Z., Alves, M.M., 2010. Methane productionfrom oleate: assessing the bioaugmentation potential ofSyntrophomonas zehnderi. Water Research 44, 4940e4947.

Demirbas, A., Sari, A., Isildak, O., 2006. Adsorptionthermodynamics of stearic acid onto bentonite. Journal ofHazardous Materials B135, 226e231.

Desbois, A.P., Smith, V.J., 2010. Antibacterial free fatty acids:activities, mechanisms of action and biotechnologicalpotential. Applied Microbiology and Biotechnology 85,1629e1642.

Diaz, G., Melis, M., Batetta, B., Angius, F., Falchi, A.M., 2008.Hydrophobic characterization of intracellular lipids in situ byNile red red/yellow emission ratio. Micron 39, 819e824.

Galbraith, H., Miller, T.B., 1973. Effect of long chain fatty acids onbacterial respiration and amino acid uptake. Journal ofApplied Microbiology 36 (4), 659e675.

Hwu, C.-S., Tseng, S.-K., Yuan, C.-Y., Kulik, Z., Lettinga, G., 1998.Biosorption of long-chain fatty acids in UASB treatmentprocess. Water Research 32 (5), 1571e1579.

Kuang, Y., Pullammanappallil, P., Lepesteur, M., Ho, G.-E., 2006.Recovery of oleate-inhibited anaerobic digestion by additionof simple substrates. Journal of Chemical Technology &Biotechnology 81, 1057e1063.

Lalman, J.A., Bagley, D.M., 2001. Anaerobic degradation andmethanogenic inhibitory effects of oleic and stearic acids.Waster Research 35, 2975e2983.

Mouneimne, A.H., Carrere, H., Bernet, J.P., Delgenes, J.P., 2004.Effect of the addition of bentonite on the anaerobicbiodegradability of solid fatty wastes. EnvironmentalTechnology 25 (4), 459e469.

Narihiro, T., Terada, T., Kikuchi, K., Iguchi, A., Ikeda, M.,Yamauchi, T., Shiraishi, K., Kamagata, Y., Nakamura, K.,Sekiguchi, Y., 2009. Comparative analysis of bacterial andarchaeal communities in methanogenic sludge granules fromupflow anaerobic sludge blanket reactors treating variousfood-processing, high-strength organic wastewaters. MicrobesEnvironment 24 (2), 88e96.

Ning, J., Kennedy, K., Fernandes, L., 1996. Biosorption of 2,4-dichorophenol by live and chemically inactivated anaerobicgranules. Water Research 30, 2039e2044.

Palatsi, J., Laureni, M., Andres, M.V., Flotats, X., Nielsen, H.B.,Angelidaki, I., 2009. Strategies for recovering inhibition causedby long chain fatty acids on anaerobic thermophilic biogasreactors. Bioresource Technology 100, 4588e4596.

Palatsi, J., Illa, J., Prenafeta-Boldu, F.X., Laureni, M., Fernandez, B.,Angelidaki, K., Flotats, X., 2010. Long-chain fatty acidsinhibition and adaptation process in anaerobic thermophilicdigestion: batch tests, microbial community structure andmathematical modelling. Bioresource Technology 101,2243e2251.

Pereira, M.A., Pires, O.C., Mota, M., Alves, M.M., 2002. Anaerobicdegradation of oleic acid by suspended sludge: identificationof palmitic acid as a key intermediate. Water Science andTechnology 45 (10), 139e144.

Pereira, M.A., Roest, K., Stams, A.J.M., Akkermans, A.D.L.,Amaral, A.L., Pons, M.-N., Ferreira, E.C., Mota, M., Alve, M.M.,2003. Image analysis, methanogenic activity measurements,and molecular biological techniques to monitor granula

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 6 8e5 2 7 85278

sludge from an EGSB reactor fed with oleic acid. Water Scienceand Technology 47 (5), 181e188.

Pereira, M.A., Pires, O.C., Mota, M., Alves, M.M., 2005. Anaerobicbiodegradation of oleic and palmitic acids: evidence of masstransfer limitations caused by long chain fatty acids.Accumulation onto the anaerobic sludge. Biotechnology andBioengineering 92 (1), 15e23.

Raposo, F., Borja, R., Sanchez, E., Martın, M.A., Martın, A.,2004. Performance and kinetic evaluation of anaerobicdigestion of two-phase olive mill effluents in reactors withsuspended and immobilized biomass. Water Research 38,2017e2026.

Rinzema, A., Boone, M., van Knippenberg, K., Lettinga, G., 1994.Bactericidal effect of long chain fatty acids in anaerobicdigestion. Water Environmental Research 66, 40e49.

Riviere, D., Desvignes, V., Pelletier, E., Chaussonnerie, S.,Guermazi, S., Weissenbach, J., Li, T., Camacho, P., Sghir, A.,2009. Towards the definition of a core of microorganismsinvolved in anaerobic digestion of sludge. ISME Journal 3 (6),700e714.

Schink, B., 1997. Energetics of syntrophic cooperation inmethanogenic degradation. Microbiology Molecular BiologyReview 61, 262e280.

Sousa, D.Z., Pereira, M.A., Stams, A.J.M., Alves, M.M., 2007.Microbial communities involved in anaerobic degradation ofunsaturated or saturated long chain fatty acids (LCFA).Applied Environmental Microbiology 73 (4), 1054e1064.

Valladao, A.B.G., Torres, A.G., Freire, D.M.G., Cammarota, M.C.,2011. Profiles of fatty acids and triacylgylcerols and theirinfluence on the anaerobic biodegradability of effluents frompoultry slaughterhouse. Bioresource Technology 102 (14),7043e7050.

Wang, Q., Garrity, G.M., Tiedje, J.M., Cole, J.R., 2007. NaıveBayesian classifier for rapid assignment of rRNA sequencesinto the new bacterial taxonomy. Applied EnvironmentalMicrobiology 73 (16), 5261e5267.

Zheng, C.J., Yoo, J.-S., Lee, T.-G., Cho, H.-Y., Kim, Y.-H.,Kim, W.-G., 2005. Fatty acid synthesis is a target forantibacterial activity of unsaturated fatty acids. FEBSLetters 579, 5157e5162.